Are we witnessing a slow rug by @Zai_org,@ZixuanLi_ ? Or has the zero flying fucks given mode been activated? GLM 5.3 and GLM 5.3 Flash have been next to unusable for at least 5 days. Errors, disconnects, agents dropping out without delivering or saving the actual work, and the list goes on. Now, the most obvious culprit is the genius decision to give away 3 weekly resets and 500M weekend tokens in one go. Their infrastructure was barely keeping up with the existing load, and they just dropped a nuke on it. The next part is even funnier — there has been, and still is, ZERO response from https://t.co/KCwelu8uva, neither on X nor Discord. Discord feels like a ghost town when it comes to mods or admins cleaning up spam from frustrated users or replying to questions. Instead, they dropped another "GIFT" of 300M FREE DAILY tokens. And shit hit the fan again. The service was barely crawling along for about 8 hours today, and then it went straight back to an endless festival of errors and timeouts. Now the sad part: It seems they simply don't care about paying customers. Whether you're on the free tier or the MAX plan, you're still getting the same useless service. Which makes me feel sorry for those who took out annual subscriptions based on their early experience, when there were next to no errors and the token allowances were insane. Cancelled.
@Zai_org @ZixuanLi_ this keeps happening for the last few weeks. Jobs getting dropped. Agents working for 1h+ and delivers zero before also getting dropped. At this point the GLM 5.3 Max is no usable at least via ZCODE app. Will try via key, but doubt that will change a thing. P.S. This is not a Free idle task.
OpenAI Dots blocked for personal Pro in EEA/UK/CH. These things are always-on, get their own cloud PC, plug into 4k apps (email, calendar, files…), do background research on your life, and keep memories even after you disconnect the app. Told ya - it's not a chatbot. That’s a personal data for sale and for grabs :D Business Premium? - Allowed. Consumer Pro? - Locked out. 🇪🇺 GDPR did its thing again.
New week, new narrative: all hail Opus 5.5, all nuke ASTRA. 🫡😂 AI Twitter has officially rotated the meta again. https://t.co/5y2GJL1jPT
Hey @0xgilbert — VOOL is getting close to release as an MIT-licensed AI runtime / agent orchestration layer, and we’re looking at integrating UsePod at the core routing level rather than as just another API endpoint. Your pricing is too good to be missed ;) Idea is: VOOL user → router discovers model + live UsePod pricing → gets x402 quote → VOOL budget/policy layer checks the cap → user’s native Solana wallet pays USDC → request is replayed with payment proof → response + receipt comes back into VOOL. No UsePod account/API key needed for the normal route. We’d also support BYOK UsePod tokens for users who already have funded balances. Feels like UsePod could become a pretty native inference/payment rail inside VOOL rather than a bolt-on provider. Would love to know if you see any holes in that flow 👀
There. Finally someone said it out loud. “Web3” was mostly just Web2 with a fucking wallet-connect button bolted on top. Seed phrase. Connect wallet. Switch chain. Bridge. Sign 4 popups. Congratulations, apparently that was the future of the internet. Dead for years — just doing the longest dead-cat bounce in tech history. And x402 “agentic payments” — cool. But for what fucking end? Agent-to-agent payments. Yes, Sherlock. Obviously agents can pay other agents. But what are they actually buying? “Services”? Which ones? I can already give my local agent one OpenRouter API key, top up one balance — even with crypto if I really want — and let it use a pile of models/providers without making the user cosplay as a blockchain engineer. That’s the actual lesson here: People don’t want crypto UX. They don’t want “Web3.” They don’t give a shit what chain your agent used. They want the fucking thing to work. Hide the rails. Keep the useful parts. Kill the ceremony. Thanks for finally explaining to the industry what normal users have been screaming for years.
